|
scran
C++ library for basic single-cell RNA-seq analyses
|
Aggregate expression values across cells. More...
#include <AggregateAcrossCells.hpp>
Classes | |
| struct | Combinations |
| Unique combinations of factors. More... | |
| struct | Defaults |
| Default parameters for aggregation. More... | |
| struct | Results |
| Container for the aggregation results. More... | |
Public Member Functions | |
| template<typename Data , typename Index , typename Factor , typename Sum , typename Detected > | |
| void | run (const tatami::Matrix< Data, Index > *input, const Factor *factor, std::vector< Sum * > sums, std::vector< Detected * > detected) |
| AggregateAcrossCells & | set_compute_sums (bool c=Defaults::compute_sums) |
| AggregateAcrossCells & | set_compute_detected (bool c=Defaults::compute_detected) |
| AggregateAcrossCells & | set_num_threads (int n=Defaults::num_threads) |
| template<typename Sum = double, typename Detected = int, typename Data , typename Index , typename Factor > | |
| Results< Sum, Detected > | run (const tatami::Matrix< Data, Index > *input, const Factor *factor) |
Static Public Member Functions | |
| template<typename Factor , typename Combined > | |
| static Combinations< Factor > | combine_factors (size_t n, std::vector< const Factor * > factors, Combined *combined) |
| template<typename Combined = int, typename Factor > | |
| static std::pair< Combinations< Factor >, std::vector< Combined > > | combine_factors (size_t n, std::vector< const Factor * > factors) |
Aggregate expression values across cells.
This class computes the sum of expression values for each grouping of cells, typically for the creation of pseudo-bulk expression profiles for cluster/sample combinations. Expression values are generally expected to be counts, though the same code can be trivially re-used to compute the average log-expression. We can also report the number of cells with detected (i.e., positive) expression values in each grouping.
|
inlinestatic |
| Factor | Factor type. |
| Combined | Type of the combined factor. May need to be different from Factor if the latter does not have enough unique levels. |
| n | Number of observations (i.e., cells). | |
| [in] | factors | Pointers to arrays of length n, each containing a different factor. |
| [out] | combined | Pointer to an array of length n, in which the combined factor is to be stored. |
Combinations object is returned containing the unique combinations of levels observed in factors. A combined factor is saved to combined; values are indices of combinations in the output Combinations object.This function compresses multiple factors into a single combined factor, which can then be used as the factor in run(). In this manner, AggregateAcrossCells can easily handle aggregation across any number of factors.
|
inlinestatic |
| Factor | Factor type. |
| Combined | Type of the combined factor. May need to be different from Factor if the latter does not have enough unique levels. |
| n | Number of observations (i.e., cells). | |
| [in] | factors | Pointers to arrays of length n, each containing a different factor. |
Combinations object, containing the unique combinations of levels observed in factors.n containing the combined factor.See the other combine_factors() method for details.
|
inline |
Compute the per-gene sum and number of cells with detected expression for each level of a grouping factor.
| Data | Type of data in the input matrix, should be numeric. |
| Index | Integer type of index in the input matrix. |
| Factor | Integer type of the factor. |
| Sum | Type of the sum, usually the same as Data. |
| Detected | Type for the number of detected cells, usually integer. |
| input | The input matrix where rows are features and columns are cells. | |
| [in] | factor | Pointer to an array of length equal to the number of columns of input, containing the factor level for each cell. All levels should be integers in $[0, N)$ where $N$ is the number of unique levels. |
| [out] | sums | Vector of length $N$ (see factor), containing pointers to arrays of length equal to the number of columns of input. These will be filled with the summed expression across all cells in the corresponding level for each gene. Alternatively, if the vector is of length 0, no sums will be computed. |
| [out] | detected | Vector of length $N$ (see factor), containing pointers to arrays of length equal to the number of columns of input. These will be filled with the number of cells with detected expression in the corresponding level for each gene. Alternatively, if the vector is of length 0, no numbers will be computed. |
|
inline |
| c | Whether to compute the sum within each factor level. |
AggregateAcrossCells object.This function only affects run() when sums and detected are not supplied as inputs.
|
inline |
| c | Whether to compute the number of detected cells within each factor level. |
AggregateAcrossCells object.This function only affects run() when sums and detected are not supplied as inputs.
|
inline |
| n | Number of threads to use. |
AggregateAcrossCells object.
|
inline |
| Sum | Type of the sum, should be numeric. |
| Detected | Type for the number of detected cells, usually integer. |
| Data | Type of data in the input matrix, should be numeric. |
| Index | Integer type of index in the input matrix. |
| Factor | Integer type of the factor. |
| input | The input matrix where rows are features and columns are cells. | |
| [in] | factor | Pointer to an array of length equal to the number of columns of input, containing the factor level for each cell. All levels should be integers in $[0, N)$ where $N$ is the number of unique levels. |
Results object is returned containing the summed expression and the number of detected cells within each factor level across all genes.This function will respect any user-supplied setting of set_compute_sums() and set_compute_detected(). If either/both are false, the corresponding statistic(s) will not be computed and the corresponding vector in Results will be empty.
1.9.8